A hierarchical latent variable model for data visualization

Christopher M. Bishop, Michael E. Tipping

    Research output: Contribution to journalArticle

    Abstract

    Visualization has proven to be a powerful and widely-applicable tool the analysis and interpretation of data. Most visualization algorithms aim to find a projection from the data space down to a two-dimensional visualization space. However, for complex data sets living in a high-dimensional space it is unlikely that a single two-dimensional projection can reveal all of the interesting structure. We therefore introduce a hierarchical visualization algorithm which allows the complete data set to be visualized at the top level, with clusters and sub-clusters of data points visualized at deeper levels. The algorithm is based on a hierarchical mixture of latent variable models, whose parameters are estimated using the expectation-maximization algorithm. We demonstrate the principle of the approach first on a toy data set, and then apply the algorithm to the visualization of a synthetic data set in 12 dimensions obtained from a simulation of multi-phase flows in oil pipelines and to data in 36 dimensions derived from satellite images.
    Original languageEnglish
    Pages (from-to)281-293
    Number of pages13
    JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
    Volume20
    Issue number3
    DOIs
    Publication statusPublished - Mar 1998

    Fingerprint

    Latent Variable Models
    Data visualization
    Data Visualization
    Hierarchical Model
    Visualization
    Multiphase flow
    Projection
    Multiphase Flow
    Satellite Images
    Expectation-maximization Algorithm
    Pipelines
    Synthetic Data
    Satellites
    High-dimensional
    Demonstrate

    Bibliographical note

    Copyright of Institute of Electrical and Electronics Engineers (IEEE)

    Keywords

    • Latent variables
    • data visualization
    • EM algorithm
    • hierarchical mixture model
    • density estimation
    • principal component analysis
    • factor analysis
    • maximum likelihood
    • clustering
    • statistics.

    Cite this

    Bishop, Christopher M. ; Tipping, Michael E. / A hierarchical latent variable model for data visualization. In: IEEE Transactions on Pattern Analysis and Machine Intelligence. 1998 ; Vol. 20, No. 3. pp. 281-293.
    @article{67f144fa5cfb4181bd52657ce2461a6b,
    title = "A hierarchical latent variable model for data visualization",
    abstract = "Visualization has proven to be a powerful and widely-applicable tool the analysis and interpretation of data. Most visualization algorithms aim to find a projection from the data space down to a two-dimensional visualization space. However, for complex data sets living in a high-dimensional space it is unlikely that a single two-dimensional projection can reveal all of the interesting structure. We therefore introduce a hierarchical visualization algorithm which allows the complete data set to be visualized at the top level, with clusters and sub-clusters of data points visualized at deeper levels. The algorithm is based on a hierarchical mixture of latent variable models, whose parameters are estimated using the expectation-maximization algorithm. We demonstrate the principle of the approach first on a toy data set, and then apply the algorithm to the visualization of a synthetic data set in 12 dimensions obtained from a simulation of multi-phase flows in oil pipelines and to data in 36 dimensions derived from satellite images.",
    keywords = "Latent variables, data visualization, EM algorithm, hierarchical mixture model, density estimation, principal component analysis, factor analysis, maximum likelihood, clustering, statistics.",
    author = "Bishop, {Christopher M.} and Tipping, {Michael E.}",
    note = "Copyright of Institute of Electrical and Electronics Engineers (IEEE)",
    year = "1998",
    month = "3",
    doi = "10.1109/34.667885",
    language = "English",
    volume = "20",
    pages = "281--293",
    journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
    issn = "0162-8828",
    publisher = "IEEE",
    number = "3",

    }

    A hierarchical latent variable model for data visualization. / Bishop, Christopher M.; Tipping, Michael E.

    In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 20, No. 3, 03.1998, p. 281-293.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - A hierarchical latent variable model for data visualization

    AU - Bishop, Christopher M.

    AU - Tipping, Michael E.

    N1 - Copyright of Institute of Electrical and Electronics Engineers (IEEE)

    PY - 1998/3

    Y1 - 1998/3

    N2 - Visualization has proven to be a powerful and widely-applicable tool the analysis and interpretation of data. Most visualization algorithms aim to find a projection from the data space down to a two-dimensional visualization space. However, for complex data sets living in a high-dimensional space it is unlikely that a single two-dimensional projection can reveal all of the interesting structure. We therefore introduce a hierarchical visualization algorithm which allows the complete data set to be visualized at the top level, with clusters and sub-clusters of data points visualized at deeper levels. The algorithm is based on a hierarchical mixture of latent variable models, whose parameters are estimated using the expectation-maximization algorithm. We demonstrate the principle of the approach first on a toy data set, and then apply the algorithm to the visualization of a synthetic data set in 12 dimensions obtained from a simulation of multi-phase flows in oil pipelines and to data in 36 dimensions derived from satellite images.

    AB - Visualization has proven to be a powerful and widely-applicable tool the analysis and interpretation of data. Most visualization algorithms aim to find a projection from the data space down to a two-dimensional visualization space. However, for complex data sets living in a high-dimensional space it is unlikely that a single two-dimensional projection can reveal all of the interesting structure. We therefore introduce a hierarchical visualization algorithm which allows the complete data set to be visualized at the top level, with clusters and sub-clusters of data points visualized at deeper levels. The algorithm is based on a hierarchical mixture of latent variable models, whose parameters are estimated using the expectation-maximization algorithm. We demonstrate the principle of the approach first on a toy data set, and then apply the algorithm to the visualization of a synthetic data set in 12 dimensions obtained from a simulation of multi-phase flows in oil pipelines and to data in 36 dimensions derived from satellite images.

    KW - Latent variables

    KW - data visualization

    KW - EM algorithm

    KW - hierarchical mixture model

    KW - density estimation

    KW - principal component analysis

    KW - factor analysis

    KW - maximum likelihood

    KW - clustering

    KW - statistics.

    UR - https://ieeexplore.ieee.org/document/667885

    U2 - 10.1109/34.667885

    DO - 10.1109/34.667885

    M3 - Article

    VL - 20

    SP - 281

    EP - 293

    JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

    JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

    SN - 0162-8828

    IS - 3

    ER -